China’s First Regulatory Framework for Virtual Companions Soon to Take Effect
Summary
China's "Interim Measures for the Administration of AI-Based Anthropomorphic Interactive Services" will take effect on July 15, 2026, establishing the country's first regulatory framework for AI virtual companions and emotional support bots. Applicable to services simulating human personality and communication for continuous emotional interaction within Mainland China, these Measures exclude general AI applications like customer service or Q&A. Providers must ensure AI systems do not generate harmful content, including that which endangers national security, promotes extremism, or induces self-harm, and must protect minors from emotional manipulation. Key obligations include establishing internal policies for algorithm and ethics review, ensuring data security and lawful training data sourcing, and encrypting user interaction data. Services must clearly label AI-generated content, provide dynamic reminders for user over-dependency, and restrict access for children under 14, implementing a "Minor Mode" with usage and spending limits. Additionally, providers must submit safety assessments for new services or when user numbers reach 1 million registered or 100,000 monthly active users.
Key takeaway
For AI developers and legal teams operating in or targeting Mainland China, these forthcoming regulations necessitate a comprehensive review of your virtual companion services. You must integrate robust content moderation, implement stringent data security protocols, and develop specific features like "Minor Mode" and dynamic dependency reminders to ensure compliance by July 15, 2026. Proactively preparing for safety assessments and algorithm reviews is crucial to avoid operational disruptions and legal penalties.
Key insights
China's new AI companion regulations prioritize content safety, data security, and user protection, especially for minors.
Principles
- AI systems must prevent content endangering national security or promoting self-harm.
- Data security and lawful sourcing are paramount for AI training and user interaction.
- Transparency about AI interaction and protection for vulnerable users are critical.
Method
Providers must perform and submit a safety assessment and report to provincial cyberspace administration for new services or significant user growth (1M registered, 100K MAU).
In practice
- Implement "Minor Mode" with usage limits and guardian alerts for children under 14.
- Deploy real-time user safety risk identification while protecting privacy.
- Encrypt user interaction data and restrict third-party sharing without consent.
Topics
- AI Regulation
- Virtual Companions
- Data Security
- Content Moderation
- Minor Protection
- China Cyberspace Administration
Best for: CTO, VP of Engineering/Data, Director of AI/ML, Legal Professional, Policy Maker, AI Ethicist
Related on AIssential
See Counsel's argued verdicts on the open AI decisions leaders are weighing →
Editorial summary, takeaway, and curation by AIssential. Original article published by Privacy & Cybersecurity Law Blog.